PHOSPHATE TRANSPORTER TRAFFIC FACILITATOR1 is a plant-specific SEC12-related protein that enables the endoplasmic reticulum exit of a high-affinity phosphate transporter in Arabidopsis

Abstract

PHOSPHATE TRANSPORTER1 (PHT1) genes encode phosphate (Pi) transporters that play a fundamental role in Pi acquisition and remobilization in plants. Mutation of the Arabidopsis thaliana PHOSPHATE TRANSPORTER TRAFFIC FACILITATOR1 (PHF1) impairs Pi transport, resulting in the constitutive expression of many Pi starvation-induced genes, increased arsenate resistance, and reduced Pi accumulation. PHF1 expression was detected in all tissues, particularly in roots, flowers, and senescing leaves, and was induced by Pi starvation, thus mimicking the expression patterns of the whole PHT1 gene family. PHF1 was localized in endoplasmic reticulum (ER), and mutation of PHF1 resulted in ER retention and reduced accumulation of the plasma membrane PHT1;1 transporter. By contrast, the PIP2A plasma membrane protein was not mislocalized, and the secretion of Pi starvation-induced RNases was not affected in the mutant. PHF1 encodes a plant-specific protein structurally related to the SEC12 proteins of the early secretory pathway. However, PHF1 lacks most of the conserved residues in SEC12 proteins essential as guanine nucleotide exchange factors. Although it functions in early secretory trafficking, PHF1 likely evolved a novel mechanism accompanying functional specialization on Pi transporters. The identification of PHF1 reveals that plants are also endowed with accessory proteins specific for selected plasma membrane proteins, allowing their exit from the ER, and that these ER exit cofactors may have a phylum-specific origin.

Figure 1.

Phenotypic Characteristics Associated with the phf1 Mutant Alleles. (A) Histochemical analysis of GUS activity driven by the IPS1:GUS reporter gene in wild-type and phf1-1 plants grown in Pi-rich (+P) or Pi-deficient (−P) medium. (B) Wild-type and phf1-1 plants after growth on medium containing 0, 60, or 1000 μM Pi (left). Details of their root hairs are shown at right. (C) Histograms of root/shoot growth ratio, cellular Pi content, and Pi and SO₄²⁻ uptake for the wild type and phf1 mutants. FW, fresh weight. (D) Wild-type, phf1-1, and phf1-2 plants after growth in Pi-deficient medium supplemented with 10 ppm arsenate. (E) Histogram showing the relative contents of different elements in the wild type and phf1 mutant alleles after growth in complete medium. In all instances, plants were grown for 12 d, except for the histochemical staining (A), root hair growth (B), and Pi and SO₄²⁻ uptake (C) experiments, in which plants were grown for 5, 5, and 8 d, respectively. Standard deviations are indicated by error bars. Statistically significant differences between the wild type and either allele (P < 0.05, according to Student's t test) are marked with asterisks.

Figure 2.

Effect of phf1 Mutations on the Expression of Pi Starvation–Responsive Genes. (A) RNA gel blot analysis of the expression of Pi-responsive genes. RNA was prepared from roots (R) and shoots (S) of the wild type and phf1-1 and phf1-2 mutants grown for 8 d in Pi-rich or Pi-deficient medium. RNA gel blots containing 15 μg of RNA were successively hybridized to the probes corresponding to the Pi starvation–responsive genes indicated. Hybridization to a probe corresponding to 18S RNA was used as a loading control. (B) Histochemical analysis of GUS activity driven by the IPS1:GUS reporter gene. Plants were grown in Pi-deficient medium with or without kinetin (left) or in low-Pi (50 μM) medium with sucrose (1 or 3%, right) or without sucrose (0%).

Figure 3.

Positional Cloning of PHF1 and the Characteristics of Its Protein. (A) Scheme of the position of PHF1 on chromosome 3 of Arabidopsis, between markers MS3-1 and AFC1. The exon structure of PHF1 is represented with boxes (dark, untranslated; light, coding region). The sequence surrounding the mutations (G-to-A transitions) in phf1-1 and phf1-2 is also shown. (B) Complementation of phf1. Histochemical analysis of IPS1:GUS activity in plants grown in Pi-rich (left) or Pi-deficient (middle) medium, and phenotype of plants grown in the presence of 10 ppm arsenate (right). The genetic constitution of the plants is as follows: wild type; phf1-1 (phf1); and phf1-1 transformed with a 6-kb genomic region corresponding to the PHF1 gene (phf1; gPHF1) or with the coding region under the control of the promoter of the 35S gene of Cauliflower mosaic virus (phf1; oPHF1). (C) Alignment of Arabidopsis PHF1 with presumed functional PHF1 homologs and with SEC12 proteins using the program T-COFFEE (Notredame et al., 2000), structural predictions using the GENESILICO metaserver (Kurowski and Bujnicki, 2003), and the SMART program (Letunic et al., 2004). The seven predicted WD repeats are indicated by gray boxes representing the predicted β-strands within each single WD (arranged in a D-to-C configuration). The dark areas in each box represent core regions predicted by the two secondary prediction methods in GENESILICO; light areas represent regions predicted by only one of the methods. The predicted transmembrane domain is indicated by a black rectangle. Colored in green are amino acid residues conserved in all PHF1 and SEC12 proteins from plants or animals used in (D); only a subset of the sequences are shown. Blue indicates amino acid residues conserved in all PHF1 proteins; pink indicates amino acid residues conserved in all plant and animal SEC12 proteins. Asterisks highlight amino acid residues conserved in all SEC12 proteins. (D) Phylogram of PHF1 and SEC12 proteins constructed with the PHYLIP software (Felsenstein, 1989). The bootstrapping value (out of 1000 samples) for each node, obtained with the same software, is shown. The proteins are as follows: AtPHF1 (Arabidopsis PHF1); LePHF1 (Lycopersicon esculentum); OsPHF1 (Oryza sativa); AtSTL2P (Arabidopsis SEC12 ortholog); OsSEC12 (O. sativa); CeSEC12 (Caenorhabditis elegans); DmSEC12 (Drosophila melanogaster); HsPREB (Homo sapiens); ScSEC12 (Saccharomyces cerevisiae); SpSTL1 (Schizosaccharomyces pombe). (E) Complementation tests with the temperature-sensitive sec12 mutant from S. cerevisiae sec12 mutant cells transformed with empty vector (p181A1NE) or with vector expressing the STL2P open reading frame or PHF1 open reading frame grown at 25 and 37°C for 3 d.

Figure 4.

Subcellular Localization of a Functional PHF1:GFP Fusion Protein. (A) Confocal laser scanning micrographs of Pi-starved root cells expressing PHF1:GFP from the PHF1 promoter. (B) Micrographs of Pi-starved leaf epidermal cells from a transgenic line harboring the 35S:PHF1:GFP construct, in which fluorescence highlights a polygonal network of ER tubules interspersed with small patches of ER lamellae, one of which is indicated with a white arrowhead. (C) to (E) Micrographs of Pi-starved leaf epidermal cells from a transgenic line harboring the 35S:PHF1:GFP construct bombarded with a construct encoding an ER-located marker (35S:DsRed2:KDEL). The images shown correspond to PHF1:GFP (C), DsRed2:KDEL (D), or an overlay of the two (E). Bars = 10 μm.

Figure 5.

Subcellular Localization of PHT1;1:GFP in Wild-Type and phf1 Plants. (A) and (B) Confocal laser scanning micrographs of Pi-starved transgenic wild-type (A) and phf1-1 (B) root cells expressing PHT1;1:GFP from the PHT1;1 promoter showing fluorescence associated with the plasma membrane and the ER, respectively. (C) and (D) Micrographs of leaf epidermal cells from transgenic wild-type (C) and phf1-1 lines (D) harboring the 35S:PHT1;1:GFP construct. In (C), fluorescence highlights the plasma membrane, and a cell junction is indicated with a white arrowhead. By contrast, in (D), a reticulate structure is very evident. Bars = 10 μm.

Figure 6.

Subcellular Localization of PIP2A Plasma Membrane Protein and RNase Secretion in the phf1 Mutant. (A) Confocal micrographs of leaf epidermal cells from transgenic wild-type and phf1-1 lines harboring the 35S:PIP2A:GFP construct showing identical fluorescence emission, associated with the plasma membrane, in both genetic backgrounds. Bars = 20 μm. (B) Secreted RNase activity from Pi-starved wild-type and mutant plants. Plants were grown for 5 d in complete liquid medium and then transferred for an additional 7 d to Pi-deficient medium. Secreted proteins were separated by native-PAGE electrophoresis and stained for RNase activity. The positions of secreted RNase1 and RNase3 are shown.

Figure 7.

Analysis of the Expression of PHT1;1:GFP in the Wild Type, the phf1 Mutant, and a PHF1-Overexpressing Transgenic Line. A transgenic line having 35S:PHT1;1:GFP at a single locus was crossed with phf1-1 and a PHF1-overexpressing transgenic line (35S:PHF1 [oPHF1]), and homozygous phf1, wild-type, and 35S:PHF1 lines containing the 35S:PHT1;1:GFP were obtained. Top, protein gel blot analysis of PHT1;1:GFP accumulation. Total protein was extracted from Pi-rich and Pi-starved plants and separated by SDS-PAGE. PHT1;1:GFP was detected by immunoblotting with anti-GFP antibody. This panel also shows Coomassie staining of the large subunit of ribulose-1,5-bis-phosphate carboxylase/oxygenase as a loading control. Bottom, RNA gel blot analysis of PHT1;1:GFP (GFP) RNA and PHF1 RNA accumulation in the Pi-starved and nonstarved plants of different genetic constitutions. RNA gel blots containing 15 μg of RNA per track were successively hybridized to the probes indicated.

Figure 8.

RNA Gel Blot Analysis of PHF1 Responsiveness to Different Developmental and Nutritional Signals, and the Effect of phr1. (A) PHF1 expression in response to developmental cues. Wild-type plants were grown for 4 weeks in soil or 2 weeks in complete medium for root isolation, and then organs were collected independently before RNA extraction: flowers (F), siliques (S), stems (St), cauline leaves (CL), nonsenescent rosette leaves (RL), senescent rosette leaves (SRL), and roots (R). (B) PHF1 responsivenes to nutritional signals. RNA was extracted from total wild-type plants grown in solid medium for 5 d and then transferred for 5 d to medium containing (control [Ct]) or lacking Pi (−P), potassium (−K), sulfur (−S), iron (−Fe), nitrogen (−N), or sucrose (−Suc). (C) PHR1 control of PHF1. RNA was extracted from wild-type, phf1-1, and phr1-1 plants grown for 8 d in complete (+P) or Pi-deficient (−P) medium. In all cases, RNA gel blots containing 15 μg of RNA per track were successively hybridized to the probes indicated.